Spaces:
Paused
Paused
File size: 9,481 Bytes
ee6e328 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 |
# Copyright 2021 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from transformers import is_vision_available
from transformers.pipelines import pipeline
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
)
from .test_pipelines_common import ANY
if is_vision_available():
from PIL import Image
else:
class Image:
@staticmethod
def open(*args, **kwargs):
pass
@is_pipeline_test
@require_vision
class ZeroShotImageClassificationPipelineTests(unittest.TestCase):
# Deactivating auto tests since we don't have a good MODEL_FOR_XX mapping,
# and only CLIP would be there for now.
# model_mapping = {CLIPConfig: CLIPModel}
# def get_test_pipeline(self, model, tokenizer, processor):
# if tokenizer is None:
# # Side effect of no Fast Tokenizer class for these model, so skipping
# # But the slow tokenizer test should still run as they're quite small
# self.skipTest("No tokenizer available")
# return
# # return None, None
# image_classifier = ZeroShotImageClassificationPipeline(
# model=model, tokenizer=tokenizer, feature_extractor=processor
# )
# # test with a raw waveform
# image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
# image2 = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
# return image_classifier, [image, image2]
# def run_pipeline_test(self, pipe, examples):
# image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
# outputs = pipe(image, candidate_labels=["A", "B"])
# self.assertEqual(outputs, {"text": ANY(str)})
# # Batching
# outputs = pipe([image] * 3, batch_size=2, candidate_labels=["A", "B"])
@require_torch
def test_small_model_pt(self):
image_classifier = pipeline(
model="hf-internal-testing/tiny-random-clip-zero-shot-image-classification",
)
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["a", "b", "c"])
# The floating scores are so close, we enter floating error approximation and the order is not guaranteed across
# python and torch versions.
self.assertIn(
nested_simplify(output),
[
[{"score": 0.333, "label": "a"}, {"score": 0.333, "label": "b"}, {"score": 0.333, "label": "c"}],
[{"score": 0.333, "label": "a"}, {"score": 0.333, "label": "c"}, {"score": 0.333, "label": "b"}],
[{"score": 0.333, "label": "b"}, {"score": 0.333, "label": "a"}, {"score": 0.333, "label": "c"}],
],
)
output = image_classifier([image] * 5, candidate_labels=["A", "B", "C"], batch_size=2)
self.assertEqual(
nested_simplify(output),
# Pipeline outputs are supposed to be deterministic and
# So we could in theory have real values "A", "B", "C" instead
# of ANY(str).
# However it seems that in this particular case, the floating
# scores are so close, we enter floating error approximation
# and the order is not guaranteed anymore with batching.
[
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
],
)
@require_tf
def test_small_model_tf(self):
image_classifier = pipeline(
model="hf-internal-testing/tiny-random-clip-zero-shot-image-classification", framework="tf"
)
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["a", "b", "c"])
self.assertEqual(
nested_simplify(output),
[{"score": 0.333, "label": "a"}, {"score": 0.333, "label": "b"}, {"score": 0.333, "label": "c"}],
)
output = image_classifier([image] * 5, candidate_labels=["A", "B", "C"], batch_size=2)
self.assertEqual(
nested_simplify(output),
# Pipeline outputs are supposed to be deterministic and
# So we could in theory have real values "A", "B", "C" instead
# of ANY(str).
# However it seems that in this particular case, the floating
# scores are so close, we enter floating error approximation
# and the order is not guaranteed anymore with batching.
[
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
],
)
@slow
@require_torch
def test_large_model_pt(self):
image_classifier = pipeline(
task="zero-shot-image-classification",
model="openai/clip-vit-base-patch32",
)
# This is an image of 2 cats with remotes and no planes
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["cat", "plane", "remote"])
self.assertEqual(
nested_simplify(output),
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
)
output = image_classifier([image] * 5, candidate_labels=["cat", "plane", "remote"], batch_size=2)
self.assertEqual(
nested_simplify(output),
[
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
]
* 5,
)
@slow
@require_tf
def test_large_model_tf(self):
image_classifier = pipeline(
task="zero-shot-image-classification", model="openai/clip-vit-base-patch32", framework="tf"
)
# This is an image of 2 cats with remotes and no planes
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["cat", "plane", "remote"])
self.assertEqual(
nested_simplify(output),
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
)
output = image_classifier([image] * 5, candidate_labels=["cat", "plane", "remote"], batch_size=2)
self.assertEqual(
nested_simplify(output),
[
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
]
* 5,
)
|